Ship cabin integrity inspection method and ship cabin integrity inspection system

Abstract

The invention provides a ship cabin integrity inspection method and a ship cabin integrity inspection system, and the ship cabin integrity inspection method comprises the steps: importing a virtual cabin model, and building a virtual-real coordinate mapping relation between the virtual cabin model and a real cabin; acquiring real pose data of the real unmanned aerial vehicle flying in the real cabin and a shot real cabin video; establishing a virtual unmanned aerial vehicle with a virtual camera in the virtual cabin model, and converting the real pose data into virtual pose data when the virtual unmanned aerial vehicle flies; controlling the virtual unmanned aerial vehicle to fly in the virtual cabin model according to the virtual pose data and recording to obtain a virtual cabin video; and carrying out picture similarity analysis on the real cabin video and the virtual cabin video under the same pose so as to evaluate the integrity of the real cabin. According to the technical scheme of the invention, a real cabin and virtual cabin combined analysis mode is adopted, so that the safety, accuracy, reliability and efficiency of the inspection process can be improved.

Description

Technical Field

[0001] This application relates to the field of shipbuilding, and more particularly, to a method and system for inspecting the integrity of ship compartments. Background Art

[0002] During the shipbuilding process, the inspection of the integrity of ship compartments is a crucial step. The integrity of ship compartments not only involves the integrity of the structure and the correct installation of equipment, but also includes the environmental control inside the compartments, the stable operation of the electrical system, and the reasonable layout of various pipelines. During the shipbuilding process, in case of missing or misinstalled structures or equipment, not only reinstallation and debugging are required, but also complex disassembly work may be involved, resulting in a significant increase in rework costs. For example, the missing installation of a key sensor may cause the entire monitoring system to malfunction, while a misinstalled valve may affect the watertightness of the entire compartment, and even cause water ingress into the compartment in extreme cases, endangering the buoyancy and stability of the ship. In addition, the misinstallation of electrical equipment may lead to short circuits or fires, seriously threatening the safety of the ship and crew. Therefore, ensuring the integrity of ship compartments is an important link that cannot be ignored in the shipbuilding and maintenance processes.

[0003] However, traditional methods for inspecting the integrity of ship compartments usually rely on manual visual inspection, that is, personnel visually inspect all positions of the compartments. This method not only has a high time cost, but also is prone to errors and omissions, resulting in poor inspection results, which in turn affects the safety of the ship and crew. In addition, traditional inspection methods also require inspectors to transfer and climb between various ship sections and units, posing a high safety risk. In recent years, with the development of unmanned equipment technology, it has gradually been applied in the inspection of large-scale industrial product manufacturing, and has shown great potential in spatial data collection, inspection, etc. However, in the field of ship compartment inspection, there is no effective method in the existing technology to accurately compare the real data collected by drones with the theoretical model, so the automated integrity inspection of large-scale intermediate ship products has not been realized, which greatly limits the reliability and efficiency of the inspection work. Summary of the Invention

[0004] The purpose of the embodiments of this application is to provide a method and system for inspecting the integrity of ship compartments. The method for inspecting the integrity of ship compartments adopts a combined analysis method of real compartments and virtual compartments, which helps to improve the safety, accuracy, and reliability of the inspection process. And because the comparison and analysis process can be easily automated, the efficiency of the inspection process is effectively improved. In addition, since the method for inspecting the integrity of ship compartments can form video data, it is convenient for post-review, archiving, and long-term maintenance and fault diagnosis.

[0005] The present application provides a first method for inspecting the integrity of a ship's cabin, including the following steps:

[0006] S101. Import a virtual cabin model and establish a mapping relationship between the virtual coordinates of the virtual cabin model and the real coordinates of the real cabin.

[0007] S102. Obtain the real pose data of a real unmanned aerial vehicle (UAV) flying in the real cabin and the real cabin video recorded by a real camera mounted on the UAV.

[0008] S103. Establish a virtual UAV with a virtual camera in the virtual cabin model. Based on the mapping relationship between the virtual and real coordinates, convert the real pose data into the virtual pose data of the virtual UAV flying in the virtual cabin model.

[0009] S104. Control the virtual UAV to fly in the virtual cabin model according to the virtual pose data and record the virtual cabin video.

[0010] S105. Perform image similarity analysis on the real cabin video and the virtual cabin video at the same pose to evaluate the integrity of the real cabin; the same pose means that after the real pose data is converted through the mapping relationship between the virtual and real coordinates, it is equal to the virtual pose data.

[0011] In an implementable solution, in step S101, a real marking point is set in the real cabin, and a real coordinate system is established with the real marking point as the origin to record the real pose data. At the same time, a corresponding virtual marking point is set at the corresponding position in the virtual cabin, and a virtual coordinate system is established with the virtual marking point as the origin to record the virtual pose data.

[0012] In an implementable solution, in step S102, the steps for obtaining the real pose data of the real UAV flying in the real cabin and the real cabin video recorded include:

[0013] S121. Control the real UAV to start flying and enter the real cabin. Use the real camera to find the real marking point, and after finding the real marking point, control the real UAV to stop at the real marking point.

[0014] S122. Control the real UAV to take off from the real marking point and fly according to a preset flight route. At the same time, start recording the real pose data and recording the real cabin video.

[0015] S123. After completing the flight process of the real UAV according to the preset flight route, stop recording the real pose data and stop recording the real cabin video, and then control the real UAV to leave the real cabin.

[0016] S124. Store the real pose data and the real cabin video in a preset location.

[0017] In an implementable solution, in step S104, the steps of controlling the virtual drone to fly in the virtual cabin model according to the virtual pose data and recording the virtual cabin video include:

[0018] S141: Place the virtual drone at the virtual marker point in the virtual cabin model;

[0019] S142: Control the virtual drone to take off from the virtual marker point and fly according to the virtual pose data, and at the same time start recording the virtual cabin video;

[0020] S143: After completing the flight process of the virtual drone according to the virtual pose data, stop recording the virtual cabin video;

[0021] S144: Store the virtual cabin video at a preset location.

[0022] In an implementable solution, a pyramid-shaped virtual frustum is set on the virtual drone. The vertex of the virtual frustum is located on the virtual camera, and the height direction of the virtual frustum is consistent with the flight direction of the virtual drone. The bottom surface of the virtual frustum is the viewing range of the virtual camera. In step S104, when the virtual drone is flying, detect and record the interference and collision situations between the virtual frustum and the models in the virtual cabin.

[0023] In an implementable solution, in step S104, after the flight process of the virtual drone ends, summarize and analyze the interference and collision situations between the virtual frustum and the models in the virtual cabin. If for any model to be detected in the virtual cabin, there is at least a certain time point when the model to be detected is completely inside the virtual frustum, the virtual camera can completely capture all the models to be detected in the virtual cabin. At this time, end step S104 and enter step S105. Otherwise, that is, the virtual camera fails to completely capture all the models to be detected in the virtual cabin. At this time, modify the flight route of the real drone according to the analysis result of the interference and collision situation, and then repeat steps S102 - S104.

[0024] In an implementable solution, in step S105, according to the preset sampling scheme, extract some picture frames from the real cabin video and the virtual cabin video respectively, and then perform picture similarity analysis on the real cabin picture frames and the virtual cabin picture frames obtained in the same pose.

[0025] In an implementable solution, the first method for inspecting the integrity of a ship's cabin further includes the following steps: S106: Send the integrity analysis result of the real cabin to the technical personnel, and the technical personnel formulate a solution and carry out repair work.

[0026] The present application also provides a second method for inspecting the integrity of a ship's cabin, including the following steps:

[0027] S201. Import a virtual cabin model and establish a mapping relationship between the virtual coordinates of the virtual cabin model and the real coordinates of the real cabin;

[0028] S202. Establish a virtual unmanned aerial vehicle (UAV) with a virtual camera in the virtual cabin model, and obtain the virtual pose data of the virtual UAV during flight in the virtual cabin and the virtual cabin video recorded by the virtual camera;

[0029] S203. Based on the mapping relationship between the virtual and real coordinates, convert the virtual pose data into the real pose data of the real UAV during flight in the real cabin model;

[0030] S204. Control the real UAV to fly in the real cabin model according to the real pose data and record the real cabin video through the real camera mounted on it;

[0031] S205. Perform image similarity analysis on the real cabin video and the virtual cabin video at the same pose to evaluate the integrity of the real cabin; the same pose means that after the real pose data is converted through the mapping relationship between the virtual and real coordinates, it is equal to the virtual pose data.

[0032] In an implementable solution, a pyramidal virtual frustum is set on the virtual UAV. The vertex of the virtual frustum is located on the virtual camera. The height direction of the virtual frustum is consistent with the flight direction of the virtual UAV, and the bottom surface of the virtual frustum is the viewing range of the virtual camera. In step S202, when the virtual UAV is flying, detect and record whether the virtual frustum interferes with or collides with the models in the virtual cabin.

[0033] In an implementable solution, in step S202, after the flight process of the virtual UAV ends, summarize and analyze the interference and collision conditions between the virtual frustum and the models in the virtual cabin. If for any model to be detected in the virtual cabin, there is at least a certain time point when the model to be detected is completely inside the virtual frustum, then the virtual camera has completely captured all the models to be detected in the virtual cabin. At this time, end step S202 and enter step S203; otherwise, the virtual camera fails to completely capture all the models to be detected in the virtual cabin. At this time, modify the flight route of the real UAV according to the analysis result of the interference and collision conditions, and then repeat step S202.

[0034] The present application also provides a ship cabin integrity inspection system, which includes an import module, an input module, a setting module, an operation module, a storage module, and an analysis module. Among them, the import module is used to import a virtual cabin model. The input module is used to input the real pose data of a real unmanned aerial vehicle (UAV) during flight in a real cabin and the real cabin video recorded. The setting module is used to establish a virtual-real coordinate mapping relationship between the virtual cabin model and the real cabin, and is also used to establish a virtual UAV with a virtual camera in the virtual cabin model. Based on the virtual-real coordinate relationship, the real pose data is converted into the virtual pose data of the virtual UAV during flight in the virtual cabin model. The operation module is used to control the virtual UAV to fly in the virtual cabin model according to the virtual pose data and record the virtual cabin video. The storage module is used to store the virtual cabin model, the real pose data, the real cabin video, the virtual pose data, and the virtual cabin video. The analysis module is used to perform picture similarity analysis on the real cabin video and the virtual cabin video in the same pose to evaluate the integrity of the real cabin, where the same pose means that after the real pose data is converted through the virtual-real coordinate mapping relationship, it is equal to the virtual pose data.

[0035] In an implementable solution, the ship cabin integrity inspection system further includes an output module, which is used to obtain the real cabin integrity analysis result from the analysis module and output the analysis result externally.

[0036] Compared with the prior art, the beneficial effects of the present application at least include:

[0037] The present application provides a ship cabin integrity inspection method, which adopts a method of combining real cabins and virtual cabins for analysis. Compared with traditional manual inspection methods, it has multiple advantages, including but not limited to: the present application uses a UAV to take pictures inside the cabin, eliminating the need to arrange personnel to enter the ship cabin, thus avoiding direct exposure of personnel to potential dangerous environments and improving the safety of the inspection process; by combining and comparing the virtual cabin video and the real cabin video for analysis, the differences between the real cabin and the design drawing can be more intuitively found, thus improving the accuracy and reliability of the inspection process; in addition, since the comparison and analysis process can be easily automated, the efficiency of the inspection process is effectively improved; video data is formed, facilitating post-event review, archiving, long-term maintenance, and fault diagnosis; the real pose data is recorded, and the virtual-real coordinate mapping relationship is used to associate the real pose data with the virtual pose data. Therefore, the real UAV and the virtual UAV do not need to fly synchronously, avoiding the problem of difficult transmission of large-capacity wireless data in a closed environment such as a ship cabin. Description of the Drawings

[0038] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the accompanying drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0039] Figure 1 It is a flowchart of the first method for inspecting the integrity of a ship's cabin according to an embodiment of the present application;

[0040] Figure 2 It is a flowchart of the second method for inspecting the integrity of a ship's cabin according to an embodiment of the present application;

[0041] Figure 3 It is a schematic diagram of a ship's cabin integrity inspection system according to an embodiment of the present application. Specific Embodiments

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Usually, the components of the embodiments of the present application described and shown in the accompanying drawings here can be arranged and designed in various different configurations.

[0043] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application to be protected, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0044] As Figure 1 shown, the present application provides a first method for inspecting the integrity of a ship's cabin, including the following steps:

[0045] S101. Import a virtual cabin model and establish a virtual-real coordinate mapping relationship between the virtual cabin model and the real cabin;

[0046] S102. Obtain the real pose data of a real unmanned aerial vehicle (UAV) flying in the real cabin and the real cabin video recorded by the real camera mounted thereon;

[0047] S103. Establish a virtual UAV with a virtual camera in the virtual cabin model, and based on the virtual-real coordinate mapping relationship, convert the real pose data into the virtual pose data of the virtual UAV flying in the virtual cabin model;

[0048] S104. Control the virtual UAV to fly in the virtual cabin model according to the virtual pose data and record to obtain a virtual cabin video;

[0049] S105. Perform image similarity analysis on the picture frames of the real cabin video and the virtual cabin video in the same pose to evaluate the integrity of the real cabin; the same pose means that after the real pose data is converted by the virtual-real coordinate mapping relationship, it is equal to the virtual pose data.

[0050] In step S101, set a real marker point in the real cabin, and establish a real coordinate system with the real marker point as the origin to record the real pose data. At the same time, set a corresponding virtual marker point at the corresponding position in the virtual cabin, and establish a virtual coordinate system with the virtual marker point as the origin to record the virtual pose data. Specifically, the real marker point can be a luminous LED lamp or a paper two-dimensional code, and there is no excessive restriction here.

[0051] In step S102, each picture frame of the real cabin video corresponds to a real pose coordinate (X1, Y1, Z1, Rx1, Ry1, Rz1), where X1, Y1, and Z1 are the position coordinates of the real UAV, and Rx1, Ry1, and Rz1 are the direction coordinates of the real UAV. All the real pose coordinates together form the real pose data. Similarly, each picture frame of the virtual cabin video corresponds to a virtual pose coordinate (X2, Y2, Z2, Rx2, Ry2, Rz2), where X2, Y2, and Z2 are the position coordinates of the virtual UAV, and Rx2, Ry2, and Rz2 are the direction coordinates of the virtual UAV. All the virtual pose coordinates together form the virtual pose data.

[0052] In addition, since most ship cabins are sealed structures and the structures are relatively complex, there may be some areas with poor lighting. Therefore, it is possible to consider mounting a lamp on the real UAV in the same direction as the real camera to ensure the shooting effect of the real cabin video, or a real camera with infrared shooting function can also be used.

[0053] The method for inspecting the integrity of a ship's cabin provided in this application adopts a method of combining real cabins and virtual cabins for analysis, and has various advantages compared with traditional manual inspection methods, including but not limited to: This application uses an unmanned aerial vehicle (UAV) to take pictures inside the cabin, eliminating the need to arrange personnel to enter the ship's cabin, thus avoiding direct exposure of personnel to potentially dangerous environments and improving the safety of the inspection process; By comparing and analyzing the virtual cabin video and the real cabin video, the differences between the real cabin and the design drawings can be more intuitively discovered, thereby improving the accuracy and reliability of the inspection process; In addition, since the comparison and analysis process can be easily automated, the efficiency of the inspection process has been effectively improved; Video data is formed, facilitating post-event review, archiving, long-term maintenance, and fault diagnosis; Real pose data is recorded, and the association between real pose data and virtual pose data is achieved by using the mapping relationship between real and virtual coordinates. Therefore, the real UAV and the virtual UAV do not need to fly synchronously, avoiding the problem of difficult transmission of large-capacity wireless data in a closed environment such as a ship's cabin.

[0054] In one embodiment, the steps for obtaining the real pose data of the real UAV during flight in the real cabin and the recorded real cabin video include:

[0055] S121. Control the real UAV to start flying and enter the real cabin. Use the real camera mounted on the real UAV to search for real marker points. After finding the real marker points, control the real UAV to stop at the real marker points.

[0056] S122. Control the real UAV to take off from the real marker points and fly along a preset flight route. At the same time, start recording the real pose data and record the real cabin video.

[0057] S123. After completing the flight process of the real UAV according to the preset flight route, stop recording the real pose data and stop recording the real cabin video. Then control the real UAV to leave the real cabin.

[0058] S124. Store the real pose data and the real cabin video in a preset location.

[0059] In one embodiment, in step S104, the steps for controlling the virtual UAV to fly in the virtual cabin model according to the virtual pose data and record the virtual cabin video include:

[0060] S141. Place the virtual UAV at the virtual marker points in the virtual cabin model.

[0061] S142. Control the virtual UAV to take off from the virtual marker points and fly according to the virtual pose data. At the same time, start recording the virtual cabin video.

[0062] After completing the flight process of the virtual UAV according to the virtual pose data, stop recording the virtual cabin video;

[0063] S144. Store the virtual cabin video at a preset location.

[0064] Before shooting the real cabin video and the virtual cabin video, the real camera and the virtual camera can be set to have the same camera parameters, including but not limited to the direction, focal length, field of view angle, lens distortion, and viewing range of the camera; at the same time, the real cabin video and the virtual cabin video can also be set to have the same video parameters, including but not limited to the duration, resolution, and frame rate of the video, so as to facilitate subsequent comparative analysis of the two. Preferably, the duration of the real cabin video / virtual cabin video can be set to 600 s, the resolution is at least 1280×720 and does not exceed 1920×1080, and the frame rate is 30 - 60 FPS. If the video resolution is too low, it may be difficult to distinguish the target to be detected in the picture, and if the video frame rate is too low, it may cause some frames to be missed during the flight of the real UAV, affecting the accuracy of the integrity analysis of the real cabin. If the video resolution or frame rate is too high, it will significantly increase the pressure on video storage and processing, and affect the efficiency of the integrity analysis of the real cabin.

[0065] In one embodiment, a pyramid-shaped virtual viewing frustum is set on the virtual UAV. The vertex of the virtual viewing frustum is located on the virtual camera, the height direction of the virtual viewing frustum is consistent with the flight direction of the virtual UAV, and the bottom surface of the virtual viewing frustum is the viewing range of the virtual camera. The parameters of the virtual viewing frustum can be set according to the parameters of the virtual camera. Therefore, the size of the virtual viewing frustum represents the viewing range of the virtual camera. When the model in the virtual cabin interferes and collides with the virtual viewing frustum, it indicates that the model to be detected enters the viewing range of the virtual camera. At the same time, since the virtual pose data and the real pose data are associated through the virtual-real coordinate mapping relationship, it is convenient to determine whether the real cabin video can capture the target to be detected in the corresponding real cabin. Therefore, in step S104, when the virtual UAV is flying, detect and record the interference and collision situation between the virtual viewing frustum and the models in the virtual cabin, and after the flight process of the virtual UAV ends, summarize and analyze the interference and collision situation between the virtual viewing frustum and the models in the virtual cabin, and the coverage of the real cabin video for the target to be detected in the real cabin can be obtained.

[0066] Specifically, for any model to be detected in the virtual cabin, if there is at least one time point when the model to be detected is completely inside the virtual frustum, it indicates that the virtual camera can completely capture all the models to be detected in the virtual cabin. At the same time, it also indicates that the real camera can completely capture all the objects to be detected in the real cabin. At this time, step S104 can be ended and step S105 can be entered.

[0067] Conversely, if one or more models to be detected do not appear in the view range of the virtual camera at all, or one or more models to be detected only partially appear in the view range of the virtual camera, it indicates that the virtual camera fails to completely capture all the models to be detected in the virtual cabin. At the same time, it also indicates that the real camera fails to completely capture all the objects to be detected in the real cabin. As a result, various problems such as misassembly, missing assembly, and damage that may exist in the undetected objects to be detected cannot be discovered, and partial damage problems existing in the objects to be detected that are not completely captured may not be discovered. At this time, the flight route of the real drone can be modified accordingly according to the analysis result of the interference and collision situation, so that the real camera can completely capture all the objects to be detected in the real cabin, and then steps S102 - S104 can be repeated.

[0068] In one embodiment, in step S105, according to a preset sampling scheme, partial picture frames are respectively extracted from the real cabin video and the virtual cabin video, and then picture similarity analysis is performed on the real cabin picture frames and the virtual cabin picture frames obtained in the same pose. For example, the picture frames can be extracted in the way of extracting one frame every other frame (that is, extracting half of the picture frames) or extracting one frame every two frames (that is, extracting one-third of the picture frames). Specifically, it can be determined according to the situation. For example, it can be determined according to the flight speed of the real drone. If the flight speed of the real drone is relatively fast, the extraction ratio can be appropriately increased to avoid missing key picture frames. By extracting partial picture frames for comparative analysis, the number of picture frames to be analyzed can be effectively reduced, thereby improving the efficiency of the inspection process.

[0069] In one embodiment, the method for inspecting the integrity of a ship cabin may further include: S106, sending the integrity analysis result of the real cabin to the technical personnel, and the technical personnel formulating a solution and carrying out repair work.

[0070] Preferably, the SURF algorithm or the SIFT algorithm can be used to identify the contours in the real cabin picture frame and the virtual cabin picture frame respectively, so as to obtain the real cabin contour map and the virtual cabin contour map. Then, the real cabin contour map and the virtual cabin contour map in the same pose can be compared and analyzed through the histogram algorithm or the hash value algorithm. If the similarity between the two is greater than or equal to 90%, it can be considered that the contents in the current real cabin contour map and the virtual cabin contour map are the same, that is, there is no problem in the current corresponding real cabin picture frame. On the contrary, if the similarity between the two is less than 90%, it is considered that there is a problem in the current corresponding real cabin picture frame. At this time, the real cabin picture frame with problems can be sorted and saved, waiting to be sent to the technical personnel for specific problem identification and solution formulation later, or the current real cabin picture frame can also be directly analyzed through the image recognition method to obtain specific problems, and then the specific problems can be sent to the technical personnel to formulate solutions.

[0071] For example, if there is a problem of missing installation of the generator fuel pump in the real cabin, the contour of the generator fuel pump can be detected in at least one virtual cabin picture frame, while the contour of the generator fuel pump cannot be detected in the corresponding real cabin picture frame in the same pose. The difference between the two can be easily analyzed through the histogram algorithm or the hash value algorithm, and then it can be judged that there is a problem of missing installation of the generator fuel pump. Similarly, if there is a situation of misinstalling the generator fuel pump as other equipment in the real cabin, or the generator fuel pump is damaged, the specific problems can also be analyzed through the foregoing method. Finally, all the problems obtained in the analysis process are sorted and summarized to form the integrity analysis result of the real cabin.

[0072] As Figure 2 shown, the present application also provides a second method for inspecting the integrity of a ship cabin, including the following steps:

[0073] S201. Import the virtual cabin model and establish the virtual-real coordinate mapping relationship between the virtual cabin model and the real cabin;

[0074] S202. Establish a virtual unmanned aerial vehicle with a virtual camera in the virtual cabin model, and obtain the virtual pose data of the virtual unmanned aerial vehicle flying in the virtual cabin and the virtual cabin video recorded by the virtual camera;

[0075] S203. Based on the virtual-real coordinate mapping relationship, convert the virtual pose data into the real pose data of the real unmanned aerial vehicle flying in the real cabin model;

[0076] S204. Control the real unmanned aerial vehicle to fly in the real cabin model according to the real pose data and record the real cabin video through the real camera mounted on it;

[0077] S205. Conduct image similarity analysis on the real cabin video and the virtual cabin video in the same pose to evaluate the integrity of the real cabin; the same pose means that after the real pose data is converted through the virtual-real coordinate mapping relationship, it is equal to the virtual pose data.

[0078] Compared with the first method for inspecting the integrity of a ship's cabin, the main difference of the second method for inspecting the integrity of a ship's cabin is that first, a test flight is carried out with a virtual cabin and a virtual drone to determine the flight route of the virtual drone that can completely capture the interior of the virtual cabin. Then, the real pose data is obtained through the virtual-real coordinate mapping relationship and the virtual pose data, so as to ensure that the real drone can completely capture the interior of the real cabin with only one flight, without the need for re-flight or supplementary flight, which helps to reduce the time and money costs of the entire process of inspecting the integrity of the ship's cabin.

[0079] Similar to the first method for inspecting the integrity of a ship's cabin, a pyramid-shaped virtual visual cone is also set on the virtual drone in the second method for inspecting the integrity of a ship's cabin. The vertex of the virtual visual cone is located on the virtual camera, the height direction of the virtual visual cone is consistent with the flight direction of the virtual drone, and the bottom surface of the virtual visual cone is the viewing range of the virtual camera.

[0080] Similarly, in step S202, when the virtual drone is flying, detect and record whether the virtual visual cone interferes and collides with the models in the virtual cabin. After the flight of the virtual drone ends, summarize and analyze the interference and collision situations between the virtual visual cone and the models in the virtual cabin.

[0081] If for any model to be detected in the virtual cabin, there is at least a certain time point when the model to be detected is completely inside the virtual visual cone, then the virtual camera has completely captured all the models to be detected in the virtual cabin. At this time, end step S202 and enter step S203; otherwise, the virtual camera fails to completely capture all the models to be detected in the virtual cabin. At this time, modify the flight route of the real drone accordingly according to the analysis result of the interference and collision situation, and then repeat step S202.

[0082] As Figure 3As shown in the figure, the present application also provides a ship cabin integrity inspection system, including an import module, an input module, a setting module, an operation module, a storage module, and an analysis module. Among them, the import module is used to import a virtual cabin model. The input module is used to input the real pose data of a real unmanned aerial vehicle (UAV) flying in a real cabin and the real cabin video recorded. The setting module is used to establish a virtual-real coordinate mapping relationship between the virtual cabin model and the real cabin, and is also used to establish a virtual UAV with a virtual camera in the virtual cabin model. Based on the virtual-real coordinate relationship, the real pose data is converted into the virtual pose data of the virtual UAV flying in the virtual cabin model. The operation module is used to control the virtual UAV to fly in the virtual cabin model according to the virtual pose data and record the virtual cabin video. The storage module is used to store the virtual cabin model, the real pose data, the real cabin video, the virtual pose data, and the virtual cabin video. The analysis module is used to perform image similarity analysis on the real cabin video and the virtual cabin video under the same pose to evaluate the integrity of the real cabin. Among them, the "same pose" means that after the real pose data is converted through the virtual-real coordinate mapping relationship, it is equal to the virtual pose data.

[0083] In addition, the ship cabin integrity inspection system may further include an output module, which is used to obtain the real cabin integrity analysis result from the analysis module and output the analysis result externally. Specifically, the output module can output the analysis result in electronic form, or print the analysis result into a paper version for technicians to view. In addition, according to the different final analysis results obtained by the analysis module, the output module can output the picture frames of the real cabin with problems for technicians to judge the problems and formulate solutions, or directly output the specific problems existing in the real cabin.

[0084] The above are only the preferred embodiments of the present application and are not used to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for inspecting the integrity of a ship cabin, characterized in that: include: S101, importing a virtual cabin model, and establishing a virtual-real coordinate mapping relationship between the virtual cabin model and a real cabin; S102, obtaining real posture data of a real drone when flying in a real cabin and a real cabin video recorded by a mounted real camera; S103, establishing a virtual drone with a virtual camera in the virtual cabin model, and converting the real posture data into virtual posture data of the virtual drone when flying in the virtual cabin model based on the virtual-real coordinate mapping relationship; S104, controlling the virtual drone to fly in the virtual cabin model according to the virtual posture data and recording a virtual cabin video; S105, performing image similarity analysis on the real cabin video and the virtual cabin video in the same posture to evaluate the integrity of the real cabin; the same posture means that the real posture data is equal to the virtual posture data after conversion through the virtual-real coordinate mapping relationship.

2. The ship cabin integrity inspection method according to claim 1, characterized in that: In step S101, a real marking point is set in the real cabin, and a real coordinate system is established with the real marking point as the origin to record the real posture data. At the same time, a corresponding virtual marking point is set at the corresponding position of the virtual cabin, and a virtual coordinate system is established with the virtual marking point as the origin to record the virtual posture data.

3. The ship cabin integrity inspection method according to claim 2, characterized in that: In step S102, the steps of acquiring the real posture data of the real drone when flying in the real cabin and the recorded real cabin video include: S121, controlling the real drone to start flying and enter the real cabin, using the real camera to find the real marking point, and after finding the real marking point, controlling the real drone to stop at the real marking point; S122, controlling the real UAV to take off from the real marking point and fly along the preset flight route, and at the same time starting to record the real posture data and the real cabin video; S123, after completing the flight process of the real drone according to the preset flight route, stop recording the real posture data and stop recording the real cabin video, and then control the real drone to leave the real cabin; S124: storing the real posture data and the real cabin video to a preset location.

4. The ship cabin integrity inspection method according to claim 2, characterized in that: In step S104, the steps of controlling the virtual drone to fly in the virtual cabin model according to the virtual posture data and recording the virtual cabin video include: S141, placing the virtual drone at a virtual marking point in the virtual cabin model; S142, controlling the virtual drone to take off from the virtual marking point and fly according to the virtual posture data, and starting to record the virtual cabin video at the same time; S143, after completing the flight process of the virtual drone according to the virtual posture data, stop recording the virtual cabin video; S144: storing the virtual cabin video to a preset location.

5. The ship cabin integrity inspection method according to claim 1, characterized in that: A pyramid-shaped virtual viewing cone is set on the virtual drone, the vertex of the virtual viewing cone is located on the virtual camera, the height direction of the virtual viewing cone is consistent with the flight direction of the virtual drone, and the bottom surface of the virtual viewing cone is the viewing range of the virtual camera; In step S104, when the virtual drone is flying, interference and collision between the virtual visual cone and the model in the virtual cabin are detected and recorded.

6. The ship cabin integrity inspection method according to claim 5, characterized in that: In step S104, after the flight process of the virtual drone is completed, the interference and collision conditions between the virtual visual cone and the models in the virtual cabin are summarized and analyzed. If for any model to be detected in the virtual cabin, there is at least a certain time point when the model to be detected is completely located inside the virtual visual cone, then the virtual camera can fully capture all models to be detected in the virtual cabin. At this time, step S104 is terminated and step S105 is entered. Otherwise, the virtual camera fails to fully capture all models to be detected in the virtual cabin. At this time, the flight path of the real drone is modified accordingly according to the results of the interference and collision analysis, and then steps S102-S104 are repeated.

7. The ship cabin integrity inspection method according to claim 1, characterized in that: In step S105, according to a preset sampling scheme, some picture frames are extracted from the real cabin video and the virtual cabin video respectively, and then picture similarity analysis is performed on the extracted real cabin picture frames and virtual cabin picture frames in the same posture.

8. The ship cabin integrity inspection method according to claim 1, characterized in that: Also includes: S106. Send the integrity analysis results of the actual cabin to the technicians, who will develop solutions and carry out repair work.

9. A method for inspecting the integrity of a ship cabin, characterized in that: include: S201, importing a virtual cabin model, and establishing a virtual-real coordinate mapping relationship between the virtual cabin model and a real cabin; S202, establishing a virtual drone with a virtual camera in the virtual cabin model, and obtaining virtual posture data of the virtual drone when flying in the virtual cabin and a virtual cabin video recorded by the virtual camera; S203, based on the virtual-real coordinate mapping relationship, converting the virtual posture data into real posture data of the real UAV when flying in the real cabin model; S204, controlling the real UAV to fly in the real cabin model according to the real posture data and obtaining the real cabin video by recording through the mounted real camera; S205, performing image similarity analysis on the real cabin video and the virtual cabin video in the same posture to evaluate the integrity of the real cabin; the same posture means that the real posture data is equal to the virtual posture data after conversion through the virtual-real coordinate mapping relationship.

10. The ship cabin integrity inspection method according to claim 9, characterized in that: A pyramid-shaped virtual viewing cone is set on the virtual drone, the vertex of the virtual viewing cone is located on the virtual camera, the height direction of the virtual viewing cone is consistent with the flight direction of the virtual drone, and the bottom surface of the virtual viewing cone is the viewing range of the virtual camera; In step S202, when the virtual drone is flying, it is detected and recorded whether the virtual visual cone interferes with or collides with the model in the virtual cabin.

11. The ship cabin integrity inspection method according to claim 9, characterized in that: In step S202, after the flight process of the virtual drone ends, the interference and collision conditions between the virtual visual cone and the models in the virtual cabin are summarized and analyzed. If for any model to be detected in the virtual cabin, there is at least a certain time point when the model to be detected is completely located inside the virtual visual cone, then the virtual camera completely captures all the models to be detected in the virtual cabin, and step S202 is terminated and step S203 is entered; otherwise, the virtual camera fails to completely capture all the models to be detected in the virtual cabin, and the flight path of the real drone is modified accordingly according to the results of the interference and collision analysis, and then step S202 is repeated.

12. A ship cabin integrity inspection system, characterized in that: include: Import module, used to import virtual cabin model; An input module is used to input the real posture data of a real drone flying in a real cabin and the recorded real cabin video; A setting module is used to establish a virtual-real coordinate mapping relationship between the virtual cabin model and the real cabin, and also to establish a virtual drone with a virtual camera in the virtual cabin model, and based on the virtual-real coordinate relationship, convert the real pose data into virtual pose data when the virtual drone flies in the virtual cabin model; An operation module is used to control the virtual drone to fly in the virtual cabin model according to the virtual posture data and record the virtual cabin video; A storage module, used for storing a virtual cabin model, real posture data, real cabin video, virtual posture data, and virtual cabin video; The analysis module is used to perform image similarity analysis on the real cabin video and the virtual cabin video in the same posture to evaluate the integrity of the real cabin; the same posture means that the real posture data is equal to the virtual posture data after conversion through the virtual-real coordinate mapping relationship.

13. The ship cabin integrity inspection system according to claim 12, characterized in that: It also includes an output module for obtaining the actual cabin integrity analysis results from the analysis module and outputting the analysis results to the outside.

Images